The stories are curated and ranked by an algorithm. We give it a list of publishers, and it does most the work, however sometimes a little human moderation is still needed for more clean, on-topic results. The newest and most popular stories appear at the top, like on Hacker News (we use a similar ranking algorithm). To measure the popularity of the stories, the algorithm uses social signals.
I think I'd echo what jcr said: it needs an additional human layer, to nail down the original source of articles. Else it's just another random aggregator site with a random set of sources, however pretty.
The algorithm is still a work in progress. We are constantly tweaking it to make it more autonomous and better at selecting stories. Thanks!